Highly accurate control of movement is fundamental to the performance of microsurgery. Vitreoretinal microsurgery in particular is among the most demanding of specialties in terms of positioning accuracy, and is likely to become more so due to the increasing interest in retinal microvascular interventions. Of similar importance when dealing with delicate tissues is precise control of applied force. The lack of it leads to complications such as iatrogenic retinal breaks and hemorrhage. As a result, surveys taken to identify directions for improvement in ophthalmic procedures have indicated that both positioning accuracy and tactile (or force) perception have high importance. In order to address the need for enhanced control of movement and force in vitreoretinal microsurgery while also maintaining the natural feel, ease of use, and direct patient contact of handheld surgical instruments, our group has developed an active handheld micromanipulator known as Micron. Micron is a fully handheld system that performs active compensation of hand tremor and other erroneous motion. We have developed novel patient-specific vision-guided position-input virtual fixtures (like a virtual stencil) that can be used with Micron for enhanced accuracy. We have also demonstrated force control of Micron. To date, however, tests of Micron (like most of the field of vitreoretinal surgical robotics) have usually been ex vivo and greatly simplified, e.g., using retina open-sky rather than an eyeball. One of the main reasons for this is the great difficulty of quantitative stereo in the nonlinear optics of the eye. We address this herein with a novel monocular structured-light approach to depth sensing. At the present stage of development of Micron, the time has now come for clinically useful interventions under realistic conditions. In this project we focus on accomplishing two which have good prospects of maximizing the benefits of Micron, and then demonstrating them in vivo. Therefore, the specific aims of this proposal are as follows: 1. To develop a complete retinal vessel cannulation system and technique using Micron that is realizable in an intact eye in vivo. This procedure will utilizea position-input virtual fixture based on microscope video and structured light projected from the tool tip. Motion scaling will help guide the surgeon to the target vessel, and avoidance zones will help prevent entry into subretinal areas. Force control will help the surgeon to avoid damaging surrounding tissue while targeting the desired vessel with a microfabricated needle. 2. To develop a complete membrane peeling system and technique using Micron that is realizable in an intact eye in vivo. This virtual fixture will enhance control of position, force, and peeling velocity. 3. To demonstrate retinal vessel cannulation and membrane peeling using Micron in a rabbit model in vivo. Performance will be evaluated in terms of tremor amplitude, applied force, and operation time, and amount of intraoperative bleeding. Postoperative SD-OCT and histology will be used to assess tissue damage.